A vital component of understanding community structure, species invasions, and the potential for ecosystem restoration is the ability to distinguish between the roles of environmental constraints versus dispersal limitation in controlling species distributions. Three main concerns are: (1) how to identify dispersal-limited species, (2) the degree to which the distributions of individual species are influenced by particular environmental constraints versus seed availability, and (3) how to locate suitable, but unoccupied sites. To address these concerns, species distribution models and experimental and observational field methods were applied in a series of three studies on annual plants in California grasslands. First, species distribution models were paired with a seeding experiment to test a novel technique for identifying dispersal-limited species. Species that distribution models suggested were niche-limited had lower fitness when planted in unoccupied sites, while species that models suggested were dispersal-limited had equivalent fitness when planted in occupied and unoccupied sites. Thus, species distribution models based on observation data successfully identified dispersal-limited species. Second, a multi-year study of the shifting distributional patterns of a single annual forb, Lupines nanus, revealed that patch locations and extents were determined by temporal variation in environmental constraints rather than by seed availability, Third, species distribution models created on multiple spatial and temporal scales, spatial data on existing species' patches, and a seeding experiment were used to identify the location of suitable unoccupied sites for Lasthenia californica and Lupines nanus. Dispersal-limited sites for Lupinus were rare and poorly predicted by environmental and/or spatial variables. Dispersal-limited sites for Lasthenia were more abundant, but generally occurred near existing patches (within 500 m). These results indicated that dispersal-limited sites may have limited use in a restoration context, since seeding of either species was found to be unlikely to establish new populations distant from existing ones. Collectively, findings from these three studies suggest that although dispersal-limited sites exist for some species, they may have only a weak influence on observed species distributions. Furthermore, in the context of ecosystem management, dispersal limitation may be of limited practical significance to restoration, emphasizing the paramount need for conservation of intact populations and ecosystems.